Recently, a new class of Micro-Electrical-Mechanical Systems (MEMS) devices have been disclosed that utilize embedded electronic charge as a means for actuation or self-generating sensors, such as those disclosed in U.S. Pat. Nos.: 6,597,560; 6,638,627; 6,688,179; 6,717,488; 6,750,590; and 6,773,488 and in US Patent Application Publication Nos.: 2002/0131228; 2002/0182091; 2003/0079543; 2003/0201784; and 2004/0023236 by way of example. Typically, charge is injected into the interface of dissimilar insulating materials by high electric field injection. Electrons “e-” are caused to tunnel into the conduction band of a material, such as silicon dioxide, from a silicon substrate via a high-applied electric field. The electrons “e-” become trapped at electronic trap sites at a composite insulator interface, such as an interface between silicon dioxide and silicon nitride. The charge remains trapped for an extremely long period of time and is therefore useful as MEMS enabling technology.
Embedded electronic charge is also useful for other macroscopic applications, such as, but not limited to, harvesting energy from the environment. These macroscopic structures include windmills that can convert the energy of wind into electrical power. However, for macroscopic structures it is impractical to embed electronic charge by tunneling into the conduction band of one member of insulating composite large structures. Typically, a suitable injecting interface, such as silicon to silicon dioxide, is not available.
One technique that has been investigated is to expose the structure to a beam of energetic particles such as an electron beam. With this technique, electrons “e-” impinge upon a surface of a composite insulating structure with sufficient energy to enter the system. These electrons “e-” are then trapped at trap sites at a dissimilar insulator interface.
Unfortunately, there is significant difficulty with this ballistic electron injection process. It is important that the energy of the arriving electrons “e-” be either below or above the range of energies where secondary electron yield is greater than unity. If the energy is within the range of greater than unity, a net positive charge can significantly affect the result. For example, positive charge within the outermost insulator layer, but close to the embedded electron charge will tend to empty the traps via internal high field, thus neutralizing the effective trapped charge. Furthermore, the simple presence of opposite sign charge in the vicinity of the trapped charge will tend to neutralize the effectiveness of the trapped charge.
Referring to FIG. 1, a graph of a secondary electron yield of silicon dioxide as a function of electron energy is shown. The data in the graph shows that the secondary electron yield is greater than unity from about 30 eV to approximately 3,800 eV. It is obvious any accelerating potential less than 30 eV does not have sufficient energy to substantially enter the system. Therefore, one must use energies greater than about 3,800 eV.
It is also desirable to create a system where the charge is as close to the surface as possible. As a result, the thickness of the outermost layer must be thin. However, as described above, the accelerating potential must be kept above the critical value of about 3,800 eV. With a thin outermost layer and the accelerating potential above about 3,800 eV, the penetration of the electrons “e-” may be too great.
Referring to FIG. 2, a graph of a Monte Carlo simulation of electron penetration into a composite 10 of a layer of silicon dioxide 12 on a layer of silicon nitride 14 on a layer of silicon dioxide 16 that is on a substrate 18 of silicon is illustrated. The layer of silicon dioxide 12, the layer of silicon nitride 14, and the layer of silicon dioxide 16 each have a thickness of about 100 nm. The thickness of the outermost layer of silicon dioxide 12 is chosen so that the average penetration depth of the arriving electrons “e-” is at the interface between the outermost layer of silicon dioxide 12 and the layer of silicon nitride 14. Unfortunately, this ballistic charge injection technique has not been shown to be effective.
Referring to FIGS. 3A and 3B, the capacitance-voltage (C-V) characteristics before and after the ballistic injection into the composite 10 are shown. The tests were performed on the composite 10 of a layer of silicon dioxide 12 on a layer of silicon nitride 14 on a layer of silicon dioxide 16 with the substrate 18 of n-type silicon with a liquid InGa top electrode. The ballistic injection parameters were 3 KeV, 100 sec., and 3,000 μC/cm2 dose.
As the graphs in FIGS. 3A and 3B show, there is severe degradation in the post-injection characteristics of the composite 10. This is presumed to be due to morphological changes creating defects. These defects apparently have a wide energy distribution and significant dipole moment. Investigations have determined poor retention time of charge for these test structures. Furthermore, the maximum-trapped charge density for these investigations is much less than that achieved using high field tunneling. Since the accelerating potential was in the range of secondary electron yield greater than unity, a slight negative shift is observed indicating the presence of positive charge.